Participatory sensing frameworks use humans and their computing devices as a large mobile sensing network. Dramatic accessibility and affordability have turned mobile devices (smartphone and tablet computers) into the most popular computational machines in the world, exceeding laptops. By the end of 2013, more than 1.5 billion people on earth will have a smartphone. Increased coverage and higher speeds of cellular networks have given these devices the power to constantly stream large amounts of data. Most mobile devices are equipped with advanced sensors such as GPS, cameras, and microphones. This expansion of smartphone numbers and power has created a sensing system capable of achieving tasks practically impossible for conventional sensing platforms. One of the advantages of participatory sensing platforms is their mobility, since human users are often in motion. This dissertation presents a set of techniques for modeling and predicting user transportation patterns from cell-phone and social media check-ins. To study large-scale transportation patterns, I created a mobile phone app, Kpark, for estimating parking lot occupancy on the UCF campus. Kpark aggregates individual user reports on parking space availability to produce a global picture across all the campus lots using crowdsourcing. An issue with crowdsourcing is the possibility of receiving inaccurate information from users, either through error or malicious motivations. One method of combating this problem is to model the trustworthiness of individual participants to use that information to selectively include or discard data. This dissertation presents a comprehensive study of the performance of different worker quality and data fusion models with plausible simulated user populations, as well as an evaluation of their performance on the real data obtained from a full release of the Kpark app on the UCF Orlando campus. To evaluate individual trust prediction methods, an algorithm selection portfolio was introduced to take advantage of the strengths of each method and maximize the overall prediction performance. Like many other crowdsourced applications, user incentivization is an important aspect of creating a successful crowdsourcing workflow. For this project a form of non-monetized incentivization called gamification was used in order to create competition among users with the aim of increasing the quantity and quality of data submitted to the project. This dissertation reports on the performance of Kpark at predicting parking occupancy, increasing user app usage, and predicting worker quality.

This dissertation undertakes theoretical and computational research to characterize and understand in detail atomic configurations and electronic structural properties of surfaces and interfaces at the nano-scale, with particular emphasis on identifying the factors that control atomic-scale diffusion and transport properties. The overarching goal is to outline, with examples, a predictive modeling procedure of stable structures of novel materials that, on the one hand, facilitates a better understanding of experimental results, and on the other hand, provide guidelines for future experimental work. The results of this dissertation are useful in future miniaturization of electronic devices, predicting and engineering functional novel nanostructures. A variety of theoretical and computational tools with different degrees of accuracy is used to study problems in different time and length scales. Interactions between the atoms are derived using both ab-initio methods based on Density Functional Theory (DFT), as well as semiempirical approaches such as those embodied in the Embedded Atom Method (EAM), depending on the scale of the problem at hand. The energetics for a variety of surface phenomena (adsorption, desorption, diffusion, and reactions) are calculated using either DFT or EAM, as feasible. For simulating dynamic processes such as diffusion of adatoms on surfaces with dislocations the Molecular Dynamics (MD) method is applied. To calculate vibrational mode frequencies, the infinitesimal displacement method is employed. The combination of non-equilibrium Green’s function (NEGF) and DFT is used to calculate electronic transport properties of molecular devices as well as interfaces and junctions.

Institutions of higher education invest a significant amount of resources in recruiting, processing, and advising new students. When students leave the institution prior to graduation, the university loses considerable revenues. Therefore, it is important for colleges and universities to refine their student recruitment and retention strategies to avoid forgone revenues by predicting which students are likely to need particular types of support services (DeBerard et al, 2004). Current models of prediction utilize extensive surveys that are impractical to administer each term, and they do not adequately identify the broad range of student persistence categories needed in order to gain a greater understanding of persistence behavior (Davidson, 2005; Porter, 2000; Tinto, 1975). This study created a linear discriminant function to predict a broad range of persistence levels of first-time freshmen students at California State University, Bakersfield (CSUB), by identifying pre-enrollment and early enrollment student variables that existed within the database of the University. This information may be used to develop support service strategies to better assist incoming students predicted to have a greater probability of not persisting.

No description available.

The impact of engineered products is a topic of increasing concern in society. The impact of a product can fall into the categories of economic, environmental, or social impact; the last category is defined as the effect of a product on the daily lives of people. Design teams lack sufficient tools to help improve the impact of products and understand the impact of products at scale in society. This dissertation aims to provide insight and methods for improving the social, environmental, and economic impact of engineered products. The majority of the research focuses on the prediction of product impacts on society, which requires a sociotechnical approach with models that contain aspects of the product and society. This begins with the introduction of an agent-based modeling approach to predict how changes to a design will ultimately impact society. Chapter 3 performs a systematic review of the literature to identify common challenges in product social impact modeling, identifies ways to mitigate the challenges, and provides a general process to create product impact models. Guidance on a general modeling process is essential to enable the widespread use of predictive impact models in engineering design. Chapter 4, provides guidance on creating sociotechnical models using primary survey data and machine learning for impact prediction using a case study of improved cookstoves in Uganda. Chapter 5 presents a method for incorporating environmental impacts, using life cycle assessment and agent-based modeling to properly scale impacts from the functional unit level to the societal level. A limitation of life cycle assessment in the early phases of product design is the difficulty of scaling impacts from the functional unit level to the population level. Using agent-based modeling together with life cycle assessment enables an understanding of the number of functional units required at the population level; allowing for the quantification of the total population-level impact. There are often trade-offs in the social, environmental, and economic sustainability space. To characterize these sustainability trade-offs, Chapter 6 illustrates the modeling of social, environmental, and economic impacts of a product and how to quantify the product sustainability trade-space. Chapter 7, presents work on identifying quantitative factors for selecting engineering global development project locations based on the potential for social impact. Finally, Chapter 8 provides the general contributions of this work, identifies limitations, and provides direction for future work. The research presented in this dissertation is a step toward a future where predictive modeling of the social, environmental, and economic impacts of products is commonplace in engineering design.

This dissertation develops explainable and sparse predictive models applied to two main healthcare applications: reproductive health and oncology. Through the application of advanced machine learning techniques and survival analysis, we aim to enhance predictive accuracy and provide actionable insights in these critical areas. The thesis is structured into four distinct problems, each focusing on a particular research question. The first problem concerns the prediction of the probability of conception among couples actively trying to conceive. Using self-reported health data from a North American preconception cohort study, we analyzed factors such as sociodemographics, lifestyle, medical history, diet quality, and specific male partner characteristics. Machine learning algorithms were employed to predict the probability of conception demonstrating improved discrimination and potential clinical utility. The second problem explores the application of machine learning algorithms to electronic health record (EHR) data for identifying predictor variables associated with polycystic ovarian syndrome (PCOS) diagnosis. Employing gradient boosted trees and feed-forward multilayer perceptron classifiers, we developed a scoring system that improved the model's performance, providing a valuable tool for early detection and intervention. The third problem focuses on predicting the risk of miscarriage among female participants who conceived during the study period. Utilizing both static and survival analysis, including Cox proportional hazard models, we developed predictive models to assess miscarriage risk. The study revealed that most miscarriages were due to random genetic errors during early pregnancy, indicating that miscarriage is not easily predicted based on preconception sociodemographic and lifestyle characteristics. Finally, the fourth problem focuses on the development of predictive models for managing Chronic Myeloid Leukemia (CML) patients. We developed models to predict whether patients will achieve deep molecular response (DMR) at later treatment stages and maintaining this status up to 60 months post-treatment initiation. These models offer insights into treatment effectiveness and patient management, aiming to support clinical decision-making and improve long-term patient outcomes. By emphasizing the explainability of these models, this dissertation not only aims to provide accurate predictions but also to ensure that the results are interpretable and actionable for healthcare professionals. Overall, this thesis showcases the potential of predictive modeling to improve reproductive health and oncology-related outcomes. The development and validation of various models in these contexts underscore the value of machine learning algorithms in healthcare research, analysis of epidemiologic data, and prediction of critical health events. The findings have significant implications for enhancing patient care, informing clinical practices, and guiding healthcare policy decisions.

Κεντρικό σημείο αναφοράς της διατριβής είναι η σύγχρονη πόλη, στο υπέδαφος της οποίας σώζονται αρχαιολογικά στρώματα. Η ανακάλυψή τους κάτω από τον ενεργό οικιστικό ιστό καθώς και η αρχαιολογική έρευνα που ακολουθεί θεωρούνται από πολλούς αιτία ανάσχεσης της κατασκευαστικής και αναπτυξιακής διαδικασίας. Ωστόσο, εδώ υποστηρίζεται ότι η ύπαρξη αρχαιολογικού υποστρώματος στην πόλη αποτελεί ένα από τα συγκριτικά πλεονεκτήματα της αναπτυξιακής της προοπτικής. Προς τούτο η ερευνητική μεθοδολογία χρησιμοποιεί ποιοτικά και ποσοτικά δεδομένα, ενώ ως μελέτη περίπτωσης επιλέγεται το Σχέδιο Πόλεως των Πατρών. Αρχικά η έρευνα βασίζεται στην αρχειακή και βιβλιογραφική επισκόπηση και στη διεξαγωγή δομημένων συνεντεύξεων με ειδικούς επιστήμονες. Στη συνέχεια, γίνεται χρήση της τεχνολογίας των Γεωγραφικών Συστημάτων Πληροφοριών και της Στατιστικής για τη δημιουργία βάσης δεδομένων, την ψηφιακή επεξεργασία της, την παραγωγή και δημιουργία προγνωστικών μοντέλων και την ανάδειξη της στατιστικής σχέσης της πολεοδομικής με την αρχαιολογική πληροφορία. Από τα αποτελέσματα, αποδεικνύεται ότι είναι δυνατή η κατασκευή μοντέλου πρόβλεψης της πιθανολογούμενης ύπαρξης αρχαίων σε μια πόλη, αλλά και του πιθανολογούμενου βάθους εντοπισμού τους, βασισμένη στην καταγραφή και επεξεργασία της πολεοδομικής και αρχαιολογικής πληροφορίας, που προέρχεται από τις εκσκαφές 5 συνεχόμενων ετών, ακόμα κι αν δεν γνωρίζει κανείς ή δεν λαμβάνει υπόψη τίποτε άλλο από την ιστορία της πόλης αυτής. Χρησιμοποιώντας αρχαιολογικές παραμέτρους σε συνδυασμό με πολεοδομικά δεδομένα είναι δυνατόν να κατασκευαστούν εξειδικευμένα μοντέλα, που μπορούν να αποτυπώσουν τις επιπτώσεις του αρχαιολογικού υποβάθρου μιας πόλης στις τρέχουσες λειτουργίες της και το αντίθετο. Τα αποτελέσματα αυτά μπορούν να χρησιμοποιηθούν τόσο σε επιχειρησιακό επίπεδο, στην άσκηση της αρχαιολογικής έρευνας και της παρακολούθησης της οικοδομικής δραστηριότητας στην πόλη, όσο και ως συμβολή σε μια ευρύτερη διερεύνηση για τη θέση της πολιτιστικής κληρονομιάς στη διαμόρφωση και προβολή της σύγχρονης πόλης. / The dissertation focuses on the contemporary city located on top of archaeological strata. Archaeological remains underneath, as well as their investigation, are considered by many as an obstacle towards the construction and development process. However, it is assumed here that the archaeological remains (below modern cities) consist a comparative advantage towards development. To justify this, qualitative and quantitative research methodology has been employed while the Town Plan of Patras, Greece is used as a case-study. Initially, an archive and literature survey takes place and structured interviews with field experts are conducted. Next, Geographical Information Systems and Statistics are applied for data processing and predictive modeling. Eventually, predictive models of the potential existence of archaeological sites and their expected depth are constructed, based on data from the excavations and the ground disturbance actions of 5 consecutive years. It becomes apparent that the results differ within the built and the unbuilt zones of a town. Using archaeological and urban parameters the impact of the archaeological background, over modern urban functions can be modeled and assessed. Moreover, the outcomes may be used by those involved in making and evaluating policies for the management of cultural heritage within planning.

One group of information systems that have attracted a lot of attention during the past decade are financial information systems, especially systems pertaining to financial markets and electronic trading. Delivering accurate and timely information to traders substantially increases their chances of making better trading decisions.Since the dawn of electronic exchanges the trading community has seen a proliferation of computer-based intelligence within the field, enabled by an exponential growth of processing power and storage capacity due to advancements in computer technology. The financial benefits associated with outperforming the market and gaining leverage over the competition has fueled the research of computational intelligence in financial information systems. This has resulted in a plethora of different techniques.The most prevalent techniques used within algorithmic trading today consist of various machine learning technologies, borrowed from the field of data mining. Neural networks have shown exceptional predictive capabilities time and time again.One recent machine learning technology that has shown great potential is Hierarchical Temporal Memory (HTM). It borrows concepts from neural networks, Bayesian networks and makes use of spatiotemporal clustering techniques to handle noisy inputs and to create invariant representations of patterns discovered in its input stream. In a previous paper [1], an initial study was carried-out where the predictive performance of the HTM technology was investigated within algorithmic trading of financial markets. The study showed promising results, in which the HTM-based algorithm was profitable across bullish-, bearish and horizontal market trends, yielding comparable results to its neural network benchmark. Although, the previous work lacked any attempt to produce near optimal trading models.Evolutionary optimization methods are commonly regarded as superior to alternative methods. The simplest evolutionary optimization technique is the genetic algorithm, which is based on Charles Darwin's evolutionary theory of natural selection and survival of the fittest. The genetic algorithm combines exploration and exploitation in the search for optimal models in the solution space.This paper extends the HTM-based trading algorithm, developed in the previous work, by employing the genetic algorithm as an optimization method. Once again, neural networks are used as the benchmark technology since they are by far the most prevalent modeling technique used for predicting financial markets. Predictive models were trained, validated and tested using feature vectors consisting of technical indicators, derived from the E-mini S&P 500 index futures market.The results show that the genetic algorithm succeeded in finding predictive models with good performance and generalization ability. The HTM models outperformed the neural network models, but both technologies yielded profitable results with above average accuracy. / Program: Magisterutbildning i informatik

This work aims at combining the Chaos theory postulates and Artificial Neural Networks classification and predictive capability, in the field of financial time series prediction. Chaos theory, provides valuable qualitative and quantitative tools to decide on the predictability of a chaotic system. Quantitative measurements based on Chaos theory, are used, to decide a-priori whether a time series, or a portion of a time series is predictable, while Chaos theory based qualitative tools are used to provide further observations and analysis on the predictability, in cases where measurements provide negative answers. Phase space reconstruction is achieved by time delay embedding resulting in multiple embedded vectors. The cognitive approach suggested, is inspired by the capability of some chartists to predict the direction of an index by looking at the price time series. Thus, in this work, the calculation of the embedding dimension and the separation, in Takens‘ embedding theorem for phase space reconstruction, is not limited to False Nearest Neighbor, Differential Entropy or other specific method, rather, this work is interested in all embedding dimensions and separations that are regarded as different ways of looking at a time series by different chartists, based on their expectations. Prior to the prediction, the embedded vectors of the phase space are classified with Fuzzy-ART, then, for each class a back propagation Neural Network is trained to predict the last element of each vector, whereas all previous elements of a vector are used as features.

College completion agendas necessarily presume year-to-year student persistence. Institutional efforts to retain admitted students has emerged for a variety of reasons, some intrinsic and others extrinsic. Some of these reasons include (1) financial exigency as institutions strive to retain tuition-paying students or meet prescribed enrollment and retention criteria currently used in performance funding strategies; (2) reputation enhancement as institutions attempt to ascend annual publications such as the U.S. News & World Report which rely on retention rates as one of several indicators used to measure institutional quality; (3) gaining a perceived advantage in admissions, marketing, and fundraising as persistence rates have, for better or worse, become a de facto measure of quality undergraduate programs; and (4) mission fulfillment as institutions, especially public institutions, are tasked with contributing towards broadly cast social goals such as access to education, economic competitiveness, and community development. Knowledge about forces that impact student attrition is critical to the development of preventative strategies that seek to improve student persistence rates. One such environmental force that has an impact on student persistence is financial aid and a student’s ability to pay for their college education. While research examining the impact of financial aid on student persistence has accumulated over the years, little is known about how the loss of certain types of aid, specifically, state-based merit aid, affects students once they enroll in an institution. The majority of studies about financial aid’s impact on student persistence were conducted prior to the establishment of many state-wide merit scholarship programs. Tinto’s (1975, 1986, 1993) interactional theory of student departure serves as the theoretical framework employed in this study. Tinto (1975) states that entering college students bring with them specific background characteristics and initial commitments that influence the student’s social and academic integration at the institution that, in turn, impact subsequent institutional and goal commitments and, ultimately, persistence. This study intends to examine pre- and post-matriculation data gathered through the admissions and financial aid processes to develop predictive models useful in calculating the probabilities associated with Bright Futures scholarship retention, institutional persistence after losing a Bright Futures scholarship award at the conclusion of a student’s first year of enrollment, and a student’s eligibility to recapture a Bright Futures scholarship award in their third year of enrollment. Data was collected passively from institutional databases on 2,418 students meeting the study criteria for inclusion in the model building process. Findings indicate that the models developed throughout the course of this study hold potential for informing institutional retention initiatives among Bright Futures scholarship award recipients.